The Valence-Bond Quantum Monte Carlo Method

TL;DR

The Valence-Bond Quantum Monte Carlo (VB-QMC) method combines classical valence-bond structures with quantum Monte Carlo techniques, enabling efficient and interpretable calculations of molecular ground and excited states.

Contribution

This paper introduces the VB-QMC approach, integrating valence-bond structures with QMC to improve efficiency and interpretability over traditional methods.

Findings

VB-VMC allows calculation of VB weights and resonance energies.

VB-QMC efficiently handles large systems on parallel computers.

Applications demonstrate its effectiveness for molecular ground and excited states.

Abstract

The VB-QMC method is presented in this chapter. It consists of using in quantum Monte Carlo (QMC) approaches with a wave function expressed as a usually short expansion of classical Valence-Bond (VB) structures supplemented by a Jastrow factor to account for dynamical correlation. Two variants exist: the VB-VMC (using variational Monte Carlo) and VB-DMC (using diffusion Monte Carlo) methods. QMC algorithms circumvent the notorious non-orthogonality issue of classical VB approaches, and allow highly efficient calculations on massively parallel machines. Calculation of VB weights and resonance energies are possible at the VB-VMC level, which makes VB-VMC a correlated method retaining all the interpretative capabilities of classical VB methods. Several recent applications are shown to illustrate the potential of this method as a modern alternative to classical VB methods to study ground and excited states of molecules.